Surface-modification of SnO₂ nanoparticles by incorporation of Al for the detection of combustible gases in a humid atmosphere

Abstract

Inhibition of hydroxyl poisoning of SnO₂ nanoparticles is important to develop a highly sensitive combustible gas sensor that functions in a humid atmosphere. For this purpose, we incorporated Al into SnO₂ nanoparticles (Al-doped SnO₂) by a precipitation method, and fabricated a thick-film-type sensor using a screen printing method. Bare SnO₂ nanoparticles and Al₂O₃-loaded SnO₂ nanoparticles were also prepared for comparison. The oxygen adsorption amount clearly decreased after Al doping and Al₂O₃ loading, according to temperature programmed desorption measurements. Al doping enhanced the sensor response (sensitivity) to H₂, CO and C₂H₅OH in a humid atmosphere by almost five to ten times. Al₂O₃ loading also slightly increased the sensor response to each gas in a humid atmosphere. The enhancement of the sensor response was attributed to both Al and Al₂O₃ acting as hydroxyl absorbers on the surface of the nanoparticles, thereby providing an oxygen adsorption site for surface combustion reactions in a humid atmosphere. Based on the relationship between the sensor response and C₂H₅OH concentration, it was estimated that Al-doped SnO₂ can detect less than one ppm C₂H₅OH in a humid atmosphere. Therefore, doping with Al, which protects and holds the adsorbed oxygen on the surface of the SnO₂, is important as a surface modification to obtain humidity-tolerant semiconductor gas sensors.

---

Introduction

Tin dioxide (SnO₂) is well known as a multifunctional material having wide applications such as gas sensors,^1–6 catalysts,^6–8 optoelectronic devices,^4,5,9–11 and so on.^4,5,12 In particular, SnO₂ based semiconductors offer great potential for gas sensors because of their detection accuracy and high sensitivity to combustible gases. For example, we developed a highly sensitive H₂S gas sensor using crystallite controlled SnO₂ nanoparticles¹ and a highly sensitive volatile organic compounds (VOCs) gas sensor using clustered (aggregated) Pd/SnO₂ nanoparticles.² Such SnO₂-based gas sensors allow instant gas detection because SnO₂ responds rapidly to combustible gases.¹³ Moreover, compact and low-power gas sensors have recently been investigated for battery-operated portable devices.^14,15 However, the potential of these gas sensors has often been demonstrated in a dry atmosphere because SnO₂ has a significantly low tolerance to humidity. Therefore, many researchers are interested in humidity-tolerant materials as semiconductor gas sensors for monitoring the gas components.

SnO₂-based semiconductors have been recognized as a promising material because of their sensitive response to combustible gases. However, SnO₂ is vulnerable to humidity because of its gas detection mechanism. Normally, the electric resistance of SnO₂ is strongly dependent on the presence of negatively charged adsorbed oxygen on the particle surface, as explained below.

O₂ + 2e⁻ → 2O_ad⁻ (1)

here e⁻ is a carrier electron in SnO₂ and O_ad⁻ is adsorbed oxygen trapping the electron. This reaction leads to the formation of an electron depletion layer on the particle surface and an increase in the electric resistance. Combustible gases, such as H₂, are detected by a combustion reaction with adsorbed oxygen on the SnO₂ surface. This reaction is shown in the following equation.

H₂ + O_ad⁻ → H₂O + e⁻ (2)

Consequently, the electric resistance decreases as the electron depletion layer is reduced by the release of the trapped electrons. The difference in the electric resistance is used as a signal for gas detection. However, oxygen adsorption on the SnO₂ surface is strongly disturbed by hydroxyl poisoning in a humid atmosphere.^16–19 Therefore, detection of combustible gases in a humid atmosphere is difficult. In a commercial gas sensor, the sensing layer is capped by a humidity absorber, such as activated carbon, to avoid hydroxyl poisoning. To develop reduced-size semiconductor gas sensors, it is desirable that the semiconductor materials are resistant to humidity.

Recent research has focused on protecting the sensor from deterioration due to hydroxyl poisoning. For example, it has been accepted that catalytic Pd loading on the SnO₂ particle leads to less hydroxyl poisoning.^18,19 We found that the main oxygen adsorption species of SnO₂ and Pd-loaded SnO₂ were O⁻ and O²⁻ in a humid atmosphere.²⁰ Accordingly, O²⁻ on PdO prevented hydroxyl adsorption on the SnO₂ surface. Furthermore, Kim et al. proposed that NiO, a humidity absorber, protects the SnO₂ surface from hydroxyl poisoning according to the results of their diffusion-reflectance Fourier transform IR (DRIFT) spectroscopy measurements.²¹ Moreover, Choi et al. reported that CuO also worked as a humidity absorber on the SnO₂ surface for H₂S sensing in a humid atmosphere. They proposed that CuO is effective because it has a high affinity towards water vapour.²² Therefore, protecting and holding adsorbed oxygen on the SnO₂ surface is the key to producing a high performance gas sensor in a humid atmosphere.

In this study, we attempted to enhance the sensor response to combustible gases in a humid atmosphere by using SnO₂-based semiconductor gas sensors. Aluminium ions and alumina, which have a high affinity for water and hydroxyl adsorption,^23–25 were incorporated into SnO₂ nanoparticles to protect the adsorbed oxygen. To investigate the effect of surface modification on the oxygen and hydroxyl adsorption, we prepared conventional spherical bare-SnO₂ nanoparticles, Al-doped SnO₂ nanoparticles, and Al₂O₃-loaded SnO₂ nanoparticles. Sensor responses to combustible gases H₂, CO, and C₂H₅OH in dry and humid atmospheres were investigated by analysing the electric resistance with considering their crystallite size, specific surface area, pore radius, and amount of oxygen adsorbed. It is noteworthy that Al-doped SnO₂ showed a high response to C₂H₅OH in a humid atmosphere even though its crystallite size, specific surface area, and pore radius were almost the same as bare-SnO₂. These results show that the developed semiconductor gas sensors can be used in a humid atmosphere without capping and could lead to the development of more compact sensor devices.

Experimental

Material preparation

Bare-SnO₂ nanoparticles were prepared using a hydrothermal treatment, as reported previously.² Stannic acid gel was prepared by adding an aqueous solution of SnCl₄·5H₂O (1 M) dropwise into a NH₄HCO₃ solution (1 M). The aqueous dispersion of the stannic acid gel was hydrothermally treated at 200 °C for 3 h at 10 MPa with continuous stirring at 600 rpm, after removal of Cl⁻ by centrifugation. To prepare a monodispersed solution of smaller SnO₂ nanoparticles, the pH of the solution was adjusted to 10.6. The solution was dried at 120 °C for 6 h and calcined at 600 °C for 3 h to produce a powder of bare-SnO₂ nanoparticles.

Al (0.5 mol%)-doped SnO₂ nanoparticles were prepared by a conventional precipitation method, reported previously.²⁶ SnCl₄·5H₂O and AlCl₃ were used as the starting materials. An appropriate amount of a mixed aqueous solution of SnCl₄ and AlCl₃ was added dropwise into a NH₄HCO₃ solution. After waiting for 12 h to obtain a white precipitate, the hydroxide gel was washed and the Cl⁻ was removed by centrifugation. The gel was dried at 120 °C and calcined at 600 °C for 3 h to obtain a powder of Al-doped SnO₂ nanoparticles.

For comparison with Al-doped SnO₂, we prepared Al₂O₃ (0.5 mol% Al atom)-loaded SnO₂ nanoparticles using an impregnation method.²⁶ Carrier SnO₂ nanoparticles were prepared using a conventional precipitation method, in the same manner as the Al-doped SnO₂. The stannic acid gel was prepared by adding a SnCl₄ aqueous solution (1 M) into NH₄HCO₃ solution (1 M). The precipitated white gel was dried at 120 and calcined 600 °C for 3 h to obtain the carrier SnO₂ powder, after removing Cl⁻ from the gel by centrifugation. The obtained powder was dispersed in deionized water and Al(NO₃)₃ (0.5 mol%) was added to the solution by stirring. The Al₂O₃-loaded SnO₂ nanoparticles powder was obtained by evaporating the solution at 100 °C and calcining the dried powder at 600 °C for 3 h.

Preparation of the sensor device and evaluation of the electric resistance

A screen printing method was used to mount the SnO₂-based powders on an alumina substrate (9 × 13 × 0.38 mm), which was printed with an Au comb-type electrode (180 μm line width, 90 μm distance between lines, 64 mm² sensing area). For screen printing, an SnO₂ paste was prepared by mixing the SnO₂-based powders with α-terpineol. The sensing films were deposited on the substrates and sintered at 580 °C for 3 h with flowing synthetic air. The air was controlled using a conventional gas-flow apparatus equipped with an electric furnace (Fig. 1). The flow rate of the synthetic air and sample gases was kept at 80 cm³ min⁻¹ by mass flow controllers (SEC-series; HORIBA STEC). Sample gases of H₂, CO, and C₂H₅OH in air were prepared by diluting the parent synthetic gas mixtures with synthetic air. A humid atmosphere was prepared by blowing the sample gases into pure water (Fig. 1), and humidity was determined using a commercial humidity sensor (TR-77Ui; T&D Corporation). Each sensor device was connected with a standard resistor in series. Voltage across the standard resistor was measured under an applied DC voltage of 4 V to evaluate the electric resistance of the sensors. An electric signal was acquired using an electrometer (2701; Keithley Instruments). The sensor response (S) was defined as the ratio of the electric resistance in air (R_a) to the resistance in the sample gases (R_g) (S = R_a/R_g).

Fig. 1 Schematic image of a gas flow apparatus with an electrical measurement system.

Results and discussion

Material characterization

XRD patterns of the bare-SnO₂, Al-doped SnO₂ (Al–SnO₂), and Al₂O₃-loaded SnO₂ (Al₂O₃–SnO₂) nanoparticles are shown in Fig. 2a. The obtained XRD patterns correspond well with the cassiterite structure of SnO₂ (JCPDS: 41-1445) and no impurity peaks were observed. The estimated crystallite size of each nanoparticle is shown in Table 1. The crystallite size of bare-SnO₂ and Al–SnO₂ were similar sizes, 14 and 12 nm, respectively. The average crystallite size of bare-SnO₂ prepared using the same conventional precipitation process as for Al–SnO₂, was almost 18 nm. Al doping reduces the crystallite size of SnO₂ nanoparticles. It is well accepted that metallic ion doping inhibits the crystal growth of SnO₂.^9,26 This is the slight evidence that Al is doped into the SnO₂ nanoparticles. A small peak shift of the XRD patterns would also indicate doping of other elements, as reported by Duan et al.⁹ However, no peak shifts in the XRD patterns were observed because the amount of Al doping was very low. In the case of Al₂O₃–SnO₂ nanoparticles, the crystallite size of 19 nm was larger than the other materials because the size of the carrier SnO₂ was 18 nm. TEM images of the nanoparticles are shown in Fig. 2b–d, and spherical type SnO₂ crystals were observed in each image. The Al₂O₃–SnO₂ nanoparticles were discernibly larger than the other nanoparticles. The observed crystallite size of bare-SnO₂, Al–SnO₂, and Al₂O₃–SnO₂ were 14, 13, and 22 nm, respectively. These were similar to the crystallite sizes estimated from the XRD patterns.

Fig. 2 (a) XRD patterns of nanoparticles and TEM images of (b) bare-SnO₂ nanoparticles, (c) Al-doped SnO₂ nanoparticles, and (d) Al₂O₃-loaded SnO₂ nanoparticles.

Table 1 Estimated values of average crystallite size (XRD), specific surface area (BET), peak pore radius (BJH) and O₂ desorption amount (O₂-TPD) for bare-SnO₂ nanoparticles, Al-doped SnO₂ nanoparticles, and Al₂O₃-loaded SnO₂ nanoparticles

Sample	Ave. crystallite size/nm	Specific surface area/m^2 g^−1	Peak pore radius/nm	O2 desorption
				μmol g^−1	μmol m^−2
Bare-SnO2	14	26.5	10.8	21	0.78
Al–SnO2	12	26.2	9.3	10	0.39
Al2O3–SnO2	19	15.7	14.3	8.0	0.51

---

The pore size distribution of each nanoparticles is shown in Fig. 3a, and the estimated specific surface area and peak pore radius are shown in Table 1. Bare-SnO₂ and Al–SnO₂ nanoparticles had a comparable specific surface area and peak pore radius. Al₂O₃–SnO₂ had a smaller specific surface area and a larger peak pore radius because of a larger crystallite size and particle aggregation. SEM images of each nanoparticles on the sensing layer are shown in Fig. 3b–d. The observed pore and particle size of bare-SnO₂ was comparable in size to Al–SnO₂. However, Al₂O₃–SnO₂ had larger nanoparticles and pores than bare-SnO₂. Both the bare-SnO₂ and Al–SnO₂ nanoparticles accumulated in a dense packing layer with small particles, whereas the Al₂O₃–SnO₂ nanoparticles accumulated in a loose packing layer. Thereby, we can investigate the Al doping effect without considering particle size and morphology.

Fig. 3 (a) Pore radius distributions of each nanoparticles and SEM images of (b) bare-SnO₂ nanoparticles, (c) Al-doped SnO₂ nanoparticles, and (d) Al₂O₃-loaded SnO₂ nanoparticles.

Fig. 4 shows the O₂-TPD (Temperature Programmed Desorption) results of bare-SnO₂, Al–SnO₂, and Al₂O₃–SnO₂ nanoparticles. Two large peaks were observed at 100 and 450 °C; these were attributed to desorption of adsorbed molecular oxygen (O₂⁻) and adsorbed dissociated oxygen with a negative charge (O⁻ or O²⁻), respectively.¹⁷ The adsorbed oxygen (O⁻ or O²⁻) peaked at 450 °C is attributed to the surface reaction (eqn (2)) that causes decreasing of adsorbed oxygen amount because of the H₂O emission evaluation.^17,20 We estimated the oxygen desorption amount in the range of 350–520 °C from the amount of adsorbed oxygen using Fig. 4, and listed these values in Table 1 as two different units, μmol g⁻¹ and μmol m⁻². Bare-SnO₂ desorbed almost two times more adsorbed oxygen (O⁻ or O²⁻) than Al–SnO₂ and Al₂O₃–SnO₂. In previous reports of TPD measurements, Huang et al. revealed that the oxygen desorption peak was observed at approximately 527 °C (800 K) when using γ-Al₂O₃.²⁷ Moreover, Putna et al. reported that no oxygen desorption peak was observed using α-Al₂O₃ (0001).²⁸ These results indicate that the presence of Al decreases the oxygen desorption amount because of the reduction of naked oxygen adsorption sites in SnO₂. The oxygen desorption amount per unit area, μmol m⁻², of Al–SnO₂ was smaller than Al₂O₃–SnO₂. Usually, the naked surface area of the carrier material decreases with an increase in dispersity of the doping or loading element. Thus, the Al in Al–SnO₂ was more dispersed on the surface than the Al₂O₃ in Al₂O₃–SnO₂, and the Al in Al–SnO₂ effectively disturbed oxygen adsorption on the SnO₂ surface. Therefore, Al doping and Al₂O₃ loading decreased the oxygen adsorption on the SnO₂, and Al was probably highly dispersed in the SnO₂ nanoparticles.

Fig. 4 Oxygen desorption spectra of O₂-TPD measurement using (a; black line) bare-SnO₂ nanoparticles, (b; red line) Al-doped SnO₂ nanoparticles, and (c; blue line) Al₂O₃-loaded SnO₂ nanoparticles.

The electric resistance in synthetic air

The electric resistance of SnO₂ nanoparticles is greatly influenced by the change in the oxygen adsorption amount⁵ and other element doping^26,29 and loading.^20,22,26 Fig. 5a and b show the electric resistance in synthetic air in dry and humid (45% R.H. at 25 °C) atmospheres, respectively. In a dry atmosphere, Al–SnO₂ and Al₂O₃–SnO₂ showed higher electric resistance than bare-SnO₂. In previous work, we prepared Fe-doped SnO₂ nanoparticles using the same process as with Al–SnO₂, and we confirmed that Fe doping into SnO₂ nanoparticles increases their electric resistance.²⁶ Hence, increasing the electric resistance of Al–SnO₂ was attributed to the replacement of Sn⁴⁺ with Al³⁺, leading to a decrease in the donor density. Conversely, mixing Al₂O₃ with SnO₂ prevents electron transport between the SnO₂ nanoparticles and increases the electric resistance. This occurs because Al₂O₃ is an insulating material. The amount of oxygen adsorbed by Al–SnO₂ and Al₂O₃–SnO₂ was less than the bare-SnO₂, as shown in Fig. 4. Therefore, Al doping and Al₂O₃ loading were effective at increasing electric resistance.

Fig. 5 Temperature dependence of the electric resistance in synthetic air atmosphere (a) in dry and (b) humid (45% R.H. at 25 °C), and (c) the response to humidity using () bare-SnO₂ nanoparticles, () Al-doped SnO₂ nanoparticles, and () Al₂O₃-loaded SnO₂ nanoparticles.

In a humid atmosphere, the electric resistance of each sensor decreased (Fig. 5b). It is well accepted that oxygen adsorption is disturbed by hydroxyls adsorbed on the oxygen adsorption site. Al–SnO₂ and Al₂O₃–SnO₂ nanoparticles showed a higher electric resistance than bare-SnO₂ nanoparticles, even in a humid atmosphere. The reason for the high electric resistance was caused less by oxygen adsorption and more by Al doping and Al₂O₃ loading. Sensor response to water vapour, which is the ratio of electric resistance in dry air to that in humid air, is shown in Fig. 5c. Sensor response increased with decreasing operating temperature in all samples because hydroxyl poisoning increases with decreasing temperatures. Bare-SnO₂ nanoparticles was clearly more sensitive to humidity than Al–SnO₂ and Al₂O₃–SnO₂ nanoparticles at 250 °C because there are fewer oxygen adsorption sites on Al–SnO₂ and Al₂O₃–SnO₂ than on bare-SnO₂. In addition, our SnO₂-based nanoparticles had a high affinity for hydroxyls.

The sensor response to combustible gases

The sensor response to combustible gases was examined using bare-SnO₂, Al–SnO₂, and Al₂O₃–SnO₂ nanoparticles. The sensor response relies on the surface reaction between adsorbed oxygen and combustible gases, so the amount of oxygen adsorbed strongly affect sensor response. Yamazoe et al. proposed a theoretical approach for estimating sensor response and suggested the following equation to represent the hydrogen response:³⁰

S² = (R_a/R_g)² = (3c/aN_d) × P_H2 + const. (3)

Here c = k₂/k₋₁ is the constant, k₂ is the rate constant of forward reaction of eqn (2), k₋₁ is the rate constant of the reverse reaction of eqn (1), a is the crystal radius, N_d is the donor density of SnO₂, and P_H2 is the partial pressure of hydrogen. According to this equation, reduction of donor density and crystallite size enhances sensor response to H₂. We have confirmed experimentally that the equation holds true for donor density reduction.²⁶ However, in the case of Al–SnO₂, there was a reduction in not only the donor density but also the oxygen adsorption amount. According to the equation, these two factors should have an opposite influence on the sensor response.

The temperature dependence of the sensor response to 200 ppm of H₂, 200 ppm of CO, and 100 ppm of C₂H₅OH in a dry atmosphere is shown in Fig. 6a–c, respectively. In response to H₂ and CO, the sensor response using bare-SnO₂ nanoparticles went through a maximum at 300 °C. Such volcano-shaped behaviour is typical of a SnO₂ sensor and can be explained theoretically by the surface reactions and gas diffusion of SnO₂.³¹ Al–SnO₂ and Al₂O₃–SnO₂ showed a high sensor response at 250 °C even though Al and Al₂O₃ are not strong catalytic materials for surface reactions. The surface combustion reaction makes byproducts H₂O and CO₂, and these byproducts are adsorbed on the oxygen adsorption site and inhibit oxygen re-adsorption. Accordingly, the sensor response is further increased. At an optimum temperature, the surface reaction activity, gas diffusion, and adsorption/desorption of byproducts are balanced to give the highest possible sensor response. The adsorption capability of the byproducts was probably higher for Al–SnO₂ and Al₂O₃–SnO₂ than bare-SnO₂ at lower temperatures. The maximum sensor response to H₂ and CO of bare-SnO₂ was higher than to Al–SnO₂ and Al₂O₃–SnO₂. As mentioned above, the sensor response is related not only to the oxygen adsorption amount but also to the crystallite size and donor density. In both the sensor responses of bare-SnO₂ and Al–SnO₂, the oxygen adsorption amount had more effect on the sensor response than the donor density. Thus, the reduced sensor response to Al–SnO₂ and Al₂O₃–SnO₂ was mainly attributed to the reduction of oxygen adsorption sites on the particle surface. The sensor response to C₂H₅OH, as shown in Fig. 6c, went through a maximum at 250 °C for all the nanoparticles. The response of bare-SnO₂ to C₂H₅OH was higher than that of Al–SnO₂ and Al₂O₃–SnO₂. The pore diameter of Al₂O₃–SnO₂ was larger than bare-SnO₂ (Fig. 3a), and the Al₂O₃–SnO₂ sensing layer was loosely packed. According to previous research, a porous sensing layer should give a higher sensor response to C₂H₅OH than densely packed sensing layer.^5,32,33 However, our sensor response to C₂H₅OH corresponded well with the oxygen adsorption amount. Zeng et al. suggested that C₂H₅OH rapidly reacts with adsorbed oxygen after adsorption and dissociation of the OH, as described in following equation:³³

C₂H₅OH_ad + O_ad⁻ → CH₃CHO + H₂O + e⁻ (4)

Fig. 6 The sensor response to (a) 200 ppm H₂, (b) 200 ppm CO, and (c) 100 ppm C₂H₅OH in dry atmosphere as a function of operating temperature using () bare-SnO₂ nanoparticles, () Al-doped SnO₂ nanoparticles, and () Al₂O₃-loaded SnO₂ nanoparticles.

According to the result shown in Fig. 5c, the SnO₂ surface has a high affinity to hydroxyls at 250 °C. Thus, the hydrophilic surface of SnO₂ is very effective for C₂H₅OH sensing because of a reaction (eqn (4)) occurring at low operating temperatures. Additionally, the C₂H₅OH combustion reaction consumes the adsorbed oxygen by successive reactions, as described in previous reports using Pt/Rh/SnO₂ (ref. 7) and Sc/In₂O₃.³⁴ Although they suggested various pathways for the combustion reaction, the clear root of the surface reaction on the bare-SnO₂ is still not clear. Nevertheless, such successive reactions enhance the sensor response to C₂H₅OH, and our SnO₂ sensors were more sensitive to C₂H₅OH than to H₂ and CO. Therefore, increasing the oxygen adsorption amount is important to enhance the sensor response by accelerating the combustion reaction. Additionally, the sensor response to C₂H₅OH is much higher than that of various SnO₂ based semiconductor gas sensors in the literature.^4–6,32,33

Fig. 7a–c shows the temperature dependence of the sensor response to H₂ and CO (200 ppm) and C₂H₅OH (100 ppm) in a humid atmosphere (45% R.H. at 25 °C). The sensor responses deteriorated significantly compared with that in a dry atmosphere because of hydroxyl poisoning. The sensor responses to H₂ and CO increased with temperature in all the materials. Increasing the temperature leads to desorption of hydroxyls, resulting in an increase of adsorbed oxygen. Accordingly, the surface reaction rate increases with temperature. For C₂H₅OH, the response went through a maximum at 300 °C because C₂H₅OH is reactive at a lower temperature than H₂ and CO, similar to that observed in the dry atmosphere. This result indicates that the amount of residual adsorbed oxygen at 300 °C is more effective than that at 350 °C. Amazingly, bare-SnO₂ shows almost no sensitivity to the target gases at 250 and 300 °C because of its high affinity to hydroxyls. This lack of sensitivity indicates that almost no adsorbed oxygen exists on the particle surface. Most importantly, the sensor response using Al–SnO₂ and Al₂O₃–SnO₂ nanoparticles was higher than bare-SnO₂ nanoparticles for all of the target gases. Such a high sensor response implies a large amount of oxygen adsorption. Hence, these results indicate that Al and Al₂O₃ inhibit hydroxyl poisoning on the oxygen adsorption site. This phenomenon is well explained by the humidity-interference effect and the introduction of a hydroxyl absorber on the SnO₂ nanoparticles, as described earlier for NiO-doped SnO₂.²¹ Al and Al₂O₃ act as a water vapour absorber on the SnO₂ surface. Additionally, the sensor response to H₂ and CO in a humid atmosphere using Al–SnO₂ nanoparticles is comparable with that of Pd-loaded SnO₂ (ref. 20) and NiO-doped SnO₂ (ref. 21) gas sensors, respectively. For C₂H₅OH, the sensor response using Al–SnO₂ nanoparticles is higher than that of porous SnO₂ gas sensors,^35,36 even though the humidity of these measurement is lower than our measurement. The sensor responses to C₂H₅OH using Al–SnO₂ nanoparticles were 2.7 times higher than that of Al₂O₃–SnO₂ nanoparticles at 300 °C, respectively. The crystallite size and specific surface area of Al₂O₃–SnO₂ were not suitable for obtaining a high sensor response. Additionally, we believe that Al in Al–SnO₂ is more highly dispersed in Al₂O₃ in Al₂O₃–SnO₂ because of the aggregation of Al₂O₃ (Fig. 4). Hence, highly dispersed Al effectively absorbs hydroxyls and accelerates oxygen adsorption on the SnO₂ surface. Thus, the sensor response to combustible gases in a humid atmosphere was enhanced.

Fig. 7 The sensor response to (a) 200 ppm H₂, (b) 200 ppm CO, and (c) 100 ppm C₂H₅OH in humid atmosphere (45% R.H. at 25 °C) as a function of operating temperature using () bare-SnO₂ nanoparticles, () Al-doped SnO₂ nanoparticles, and () Al₂O₃-loaded SnO₂ nanoparticles.

Fig. 8 shows the sensor response to C₂H₅OH as a function of C₂H₅OH concentration at 300 °C in a humid atmosphere (45% R.H. at 25 °C) using bare-SnO₂, Al–SnO₂, and Al₂O₃–SnO₂ nanoparticles. The sensor responses were clearly proportional to the C₂H₅OH concentration. According to the extrapolation line, it can be expected that Al–SnO₂ nanoparticles detect less than one ppm C₂H₅OH in a humid atmosphere. We believe that optimizing the Al doping amount would give SnO₂-based materials a more sensitive surface in a humid atmosphere.

Fig. 8 The C₂H₅OH concentration dependence of the sensor response at 300 °C in humid atmosphere, using () bare-SnO₂ nanoparticles, () Al-doped SnO₂ nanoparticles, and () Al₂O₃-loaded SnO₂ nanoparticles.

Conclusions

We have successfully developed highly sensitive, surface-modified SnO₂ nanoparticles incorporating Al and Al₂O₃ that demonstrate resistance to humidity. Spherical bare-SnO₂, Al–SnO₂, and Al₂O₃–SnO₂ nanoparticles were prepared using conventional methods. XRD, TEM, surface analysis, and SEM suggest that the bare-SnO₂ nanoparticle and Al–SnO₂ nanoparticle are similar in size while the Al₂O₃–SnO₂ is larger. The amount of oxygen adsorbed by bare-SnO₂ was clearly greater than that adsorbed by Al–SnO₂ and Al₂O₃–SnO₂, as confirmed by O₂-TPD measurement. Bare-SnO₂ adsorbed more oxygen because it has a high affinity to oxygen ions. In addition, the dispersity of Al in Al–SnO₂ was higher than that of Al₂O₃ in Al₂O₃–SnO₂. The sensor responses to combustible gases using Al–SnO₂ and Al₂O₃–SnO₂ were lower than bare-SnO₂ in a dry atmosphere because incorporating Al reduces the combustion reaction rate with adsorbed oxygen. Conversely, in a humid atmosphere, the response to combustible gases using Al–SnO₂ was approximately five to ten times larger than bare-SnO₂. Al₂O₃–SnO₂ also showed a higher sensor response to combustible gases than bare-SnO₂. These results indicate that Al and Al₂O₃ inhibit hydroxyl poisoning of SnO₂ nanoparticles by working as hydroxyl absorbers and producing an oxygen adsorption site around them. The Al doping process dispersed Al in SnO₂ nanoparticles discretely, leading to an effective surface for combustion reactions in a humid atmosphere. Additionally, it was expected that Al–SnO₂ can detect C₂H₅OH at ppb levels in a humid atmosphere. Therefore, Al doping offers a better surface for combustible gas sensors in a humid atmosphere. These results suggest surface modification as a possible improvement for material design. We believe that further modification of Al–SnO₂, such as morphological regulation and catalytic addition, would create attractive materials for humidity-tolerant semiconductor gas sensors. Therefore, further development of these materials will enable development of the high performance gas sensors with resistance to poisoning gases and more compact gas sensors than are currently available.

Acknowledgements

This work was supported in part by the FIST Foundation for Interaction in Science & Technology, Japan.

Notes and references

1. T. Kida, S. Fujiyama, K. Suematsu, M. Yuasa and K. Shimaoe, Pore and Particle Size Control of Gas Sensing Films Using SnO₂ Nanoparticles Synthesized by Seed-Mediated Growth: Design of Highly Sensitive Gas Sensors, J. Phys. Chem. C, 2013, 117, 17574–17582.
2. K. Suematsu, Y. Shin, Z. Hua, K. Yoshida, T. Kida and K. Shimanoe, Nanoparticle Cluster Gas Sensor: Controlled Clustering of SnO₂ Nanoparticles for Highly Sensitive Toluene Detection, ACS Appl. Mater. Interfaces, 2014, 6, 5319–5326.
3. D. Flak, A. Braun, B. S. Mun, J. B. Park, M. Parlinska-Wojtan, T. Graule and M. Rekas, Spectroscopic Assessment of the Role of Hydrogen in Suface Defects, in the Electronik Structure and Transport Properties of TiO₂, ZnO and SnO₂ Nanoparticles, Phys. Chem. Chem. Phys., 2013, 15, 1417–1430.
4. A. Kar and A. Patra, Recent Development of Core-Shell SnO₂ Nanostructures and Their Potential Applications, J. Mater. Chem. C, 2014, 2, 6707–6722.
5. Y. Liu, Y. Jiao, Z. Zhang, F. Qu, A. Umar and X. Wu, Hierarchical SnO₂ Nanostructures Made of Intermingled Ultrathin Nanosheets for Environmental Remediation, Smart Gas Sensor, and Supercapacitor Applications, ACS Appl. Mater. Interfaces, 2014, 6, 2174–2184.
6. V. V. Kovalenko, M. N. Rumyantseva, A. M. Gaskov, E. V. Makshina, V. V. Yushchenko, I. I. Ivanova, A. Ponzoni, G. Faglia and E. Comini, SnO₂/Fe₂O₃ Nanocomposites: Ethanol-Sensing Performance and Catalytic Activity for Oxidation of Ethanol, Inorg. Mater., 2006, 42, 1088–1093.
7. A. Kowal, M. Li, M. Shao, K. Sasaki, M. B. Vukmirovic, J. Zhang, N. S. Marinkovic, P. Liu, A. I. Frenkel and R. R. Adzic, Ternary Pt/Rh/SnO₂ electrocatalysts for oxidizing ethanol to CO₂, Nat. Mater., 2009, 8, 325–330.
8. N. Kamiuchi, T. Mitsui, N. Yamaguchi, H. Muroyama, T. Matsui, R. Kikuchi and K. Eguchi, Activation of Pt/SnO₂ Catalyst for Catalytic Oxidation of Volatile Organic Compounds, Catal. Today, 2010, 157, 415–419.
9. Y. Duan, J. Zheeng, N. Fu, Y. Fang, T. Liu, Q. Zhang, X. Zhou, Y. Lin and F. Pan, Enhancing the Performance of Dye-Sensitized Solar Cells: Doping SnO₂ Photoanodes with Al to Simultaneously Improve Conduction Band and Electron Lifetime, J. Mater. Chem. A, 2015, 3, 3066–3073.
10. T. Jia, W. Wang, F. Long, Z. Fu, H. Wang and Q. Zhang, Synthesis, Characterization, and Photocatalytic Activity of Zn-Doped SnO₂ Hierarchical Architectures Assembled by Nanocones, J. Phys. Chem. C, 2009, 113, 9071–9077.
11. X. Duo, D. Sabba, N. Mathews, L. Helena, Y. M. Lam and S. Mhaisalkar, Hydrothermal Synthesis of High Electron Mobility Zn-doped SnO₂ Nanoflowers as Photoanode Material for Efficient Dye-Sensitized Solar Cells, Chem. Mater., 2011, 23, 3938–3945.
12. J. Zhu, Z. Lu, S. T. Aruna, D. Aurbach and A. Gedanken, Sonochemical Synthesis of SnO₂ Nanoparticles and Their Preliminary Study as Li Insertion Electrodes, Chem. Mater., 2000, 12, 2557–2566.
13. T. Kida, T. Kuroiwa, M. Yuasa, K. Shimanoe and N. Yamazoe, Study on the Response and Recovery Properties of Semiconductor Gas Sensors Using a High-Speed Gas-Switching System, Sens. Actuators, B, 2008, 134, 928–933.
14. K. Suematsu, Y. Shin, N. Ma, T. Oyama, M. Sasaki, M. Yuasa, T. Kida and K. Shimanoe, Pulse-Driven micro Gas Sensor Fitted with Clustered Pd/SnO₂ Nanoparticles, Anal. Chem., 2015, 87, 8407–8415.
15. R. Triantafyllopoulou and C. Tsamis, Detection of CO and NO Using Low Power Metal Oxide Sensors, Phys. Status Solidi, 2008, 205, 2643–2646.
16. D. Koziej, N. Barsan, U. Weimar, J. Szuber, K. Shimanoe and N. Yamazoe, Water-Oxygen Interplay on Tin Dioxide Surface: Inplication on Gas Sensing, Chem. Phys. Lett., 2005, 410, 321–323.
17. N. Yamazoe, J. Fuchigami, M. Kishikawa and T. Seiyama, Interaction of Tin Oxide Surface with O₂, H₂O and H₂, Surf. Sci., 1979, 86, 335–344.
18. R. G. Pavelko, H. Daly, C. Hardacre, A. A. Vasiliev and E. Llobet, Interaction of Water, Hydrogen and Their Mixtures with SnO₂ Based Materials: the Role of Surface Hydroxyl Groups in Detection Mecahnisms, Phys. Chem. Chem. Phys., 2010, 12, 2639–2647.
19. G. Korotchenkov, V. Brynzari and S. Dmitriev, Electrical Behavior of SnO₂ Thin Films in Humid Atmosphere, Sens. Actuators, B, 1997, 54, 197–201.
20. N. Ma, K. Suematsu, M. Yuasa, T. Kida and K. Shimanoe, Effect of Water Vapor on Pd-Loaded SnO₂ Nanoparticles Gas Sensor, ACS Appl. Mater. Interfaces, 2015, 7, 5863–5869.
21. H.-R. Kim, A. Haensch, I.-D. Kim, N. Barsan, U. Weimar and J.-H. Lee, The Role of NiO Doping in Reducing the Impact of Humidity on the Performance of SnO₂-Based Gas Sensors: Synthesis Strategies, and Phenomenological and Spectroscopic Studies, Adv. Funct. Mater., 2011, 21, 456–4463.
22. K.-I. Choi, H.-J. Kim, Y.-C. Lang and J.-H. Lee, Ultraselective and ultrasensitive detection of H₂S in highly humid atmosphere using CuO-loaded SnO₂ hollow spheres for real-time diagnosis of halitosis, Sens. Actuators, B, 2014, 194, 371–376.
23. J. B. Peri, A Model for the Surface of γ-Alumina, J. Phys. Chem., 1965, 69, 220–230.
24. J. Szanyi, J. H. Kwak, R. J. Chimentao and C. H. F. Peden, Effect of H₂O on the Adsorption of NO₂ on γ-Al₂O₃: an in Situ FTIR/MS Study, J. Phys. Chem., 2007, 111, 2661–2669.
25. T. Shirai, J. W. Li, K. Matsumaru, C. Ishizaki and K. Ishizaki, Surface Hydration States of Commercial High Purity α-Al₂O₃ Powders Evaluated by Temperature Programmed Desorption Mass Spectrometry and Diffuse Reflectance Infrared Fourier Transform Spectroscopy, Sci. Technol. Adv. Mater., 2005, 6, 123–128.
26. K. Suematsu, M. Yuasa, T. Kida, N. Yamazoe and K. Shimanoe, Effects of Crystallite Size and Donor Density on the Sensor Response of SnO₂ Nano-Particles in the State of Volume Depletion, J. Electrochem. Soc., 2012, 159, J136–J141.
27. S. J. Huang, A. B. Walters and M. A. Vannice, TPD, TPR and DRIFT Studies of Adsorption and reduction of NO on La₂O₃ dispersed on Al₂O₃, Appl. Catal., B, 2000, 26, 101–118.
28. E. S. Putna, J. M. Vohs and R. J. Gorte, Evidence for Weakly Bound Oxygen on Ceria Films, J. Phys. Chem., 1996, 100, 17862–17865.
29. S. Singkammo, A. Wisitsoraat, C. Sriprachuabwong, A. Tuantranont, S. Phanichphant and C. Liewhiran, Electrolytically Exfoliated Graphene-Loaded Flame-Made Ni-Doped SnO₂ Composite Film for Acetone Sensing, ACS Appl. Mater. Interfaces, 2015, 7, 3077–3092.
30. N. Yamazoe and K. Shimanoe, Role of Shape and Size of Component Crystals in Semiconductor Gas Sensors II. Response to NO₂ and H₂, J. Electrochem. Soc., 2008, 155, J93–J98.
31. G. Sakai, N. Matsunaga, K. Shimanoe and N. Yamazoe, Theory of Gas-Diffusion Controlled Sensitivity for Thin Film Semiconductor Gas Sensor, Sens. Actuators, B, 2001, 80, 125–131.
32. J. Xu, Y. S. Li, H. T. Huang, Y. G. Zhu, Z. R. Wang, Z. Xie, X. Wang, D. Chen and G. Shen, Synthesis, Characterizations and Improved Gas-Sensing Performance of SnO₂ Nanospike Arrays, J. Mater. Chem., 2011, 21, 19086–19092.
33. W. Zeng, T. Li, T. Li, J. Hao and Y. Li, Template-Free Synthesis of Highly Ethanol-Response Hollow SnO₂ Spheres Using Hydrothermal Process, J. Mater. Sci.: Mater. Electron., 2015, 26, 1192–1197.
34. M. Iwamoto, M. Tanaka, S. Hirakawa, S. Mizuno and M. Kurosawa, Pulse and IR Study on the Reaction Pathways for the Conversion of Ethanol to Propene over Scandium-Loaded Indium Oxide Catalysts, ACS Catal., 2014, 4, 3463–3469.
35. J. Zhang, J. Guo, H. Xu and B. Gao, Reactive-Template Fabrication of Porous SnO₂ Nanotubes and Their Remarkable Gas-Sensing Performance, ACS Appl. Mater. Interfaces, 2013, 5, 7893–7898.
36. H. Zhang, W. Zeng, Y. Zhang, Y. Li, B. Miao, W. Chen and X. Peng, Synthesis and Gas Sensing Properties of Novel SnO₂ Nanorods, J. Mater. Sci.: Mater. Electron., 2014, 25, 5006–5012.

---

Footnotes

† Current address: Chemical and Texture Research Institute, Fukuoka Industrial Technology Center, Chikushino, Fukuoka, 818-8540, Japan.

‡ Current address: Department of Biological and Environmental Chemistry, School of Humanity-Oriented Science and Engineering, Kinki University, Iizuka, Fukuoka 820-8555, Japan.

---